I will reiterate that the standard statistical method for combining dissimilar quantities into a "score" is the geometric mean, and that is what I think we should use.  Beyond that, I simply don't have the time or energy to keep arguing about it.

Punchy


I will reiterate that the standard statistical method for combining dissimilar quantities into a "score" is the geometric mean, and that is what I think we should use.  Beyond that, I simply don't have the time or energy to keep arguing about it.

 
See #588.

I don't have the time to argue either. So, please do as you wish.

Now now all.... You have to admit, regardless of personal angst... This past few days has had a lot of kick butt efforts from all. 
 
As I always try to help my employees understand... You cannot make bread without mixing the ingredients, and it has to get hot before it can cool...

troy8d

devdog51


Alright Gryphon since the CCC is your baby, Im going to pose Adak's questions to you directly:

** Removed** 

Personal Attacks - Personal attacks often lead to the trading of insults, and can throw the discussion of a legitimate topic off-track.  Personal attacks can be defined as issuing a single or repeated personal attack or attacks aimed at another member, rather than at their opinions or ideas.  Criticizing a member's post is not considered a personal attack, but using terms such as "child," "idiot," "fanboy," for example, or any other derogatory term designed to discredit a member, is not permitted.  Furthermore, comments of a racist or sexist nature, as well as derogatory comments about national origin or sexual orientation, will be dealt with harshly as in a ban from the forums.

Sorry troy I forgot. I didnt mean to exclude you, I was just caught up in getting ready for work, cleaning my house and trying to help resolve issues.

theGryphon


I don't have the time to argue either. So, please do as you wish.

 
I do believe they do the same thing, one is just a much simpler approach than the other. 

Listen im not trying to to offend anyone here, Im not trying to push for anything that no one wants, but we keep going around in circles. Here's what the argument keeps sounding like. "Well we need to decide on the system, but first we need to decide on the date, but first we need to decide on the system, but we need to decide on the date so we can figure out if we have time to implement the system. All right, here's one thing we can decide now. Since the polls suggest everyone keeps their own names is the popular idea, is that what EVGA is in agreement on? And Adak had a question before we decide the timing:  who is going to set and monitor the web site this year, and how much lead time will he/she need? Lets take care of this and then move on to the next matter. Im getting tired of reading in circles.

great job so far guys....it's giving me time to get stuff done
still have some running around to do this week.
 
the old man's not doing too bad, but I have bills to pay (I have to take over now), things to get,shopping, home care equipment, appointments to make or move, etc...
 
all the things one has to do when you have elderly parents to take care of
 
So what the current format/formula/date?
 
 
 
 
 

Xavier Zepherious


great job so far guys....it's giving me time to get stuff done
still have some running around to do this week.

the old man's not doing too bad, but I have bills to pay (I have to take over now), things to get,shopping, home care equipment, appointments to make or move, etc...

all the things one has to do when you have elderly parents to take care of

So what the current format/formula/date?

The current formula that Zodac likes and that troy created a few pages back. The CCC. Adak is against it because it favors the big teams and needs tweaked some more. As for date thats still going round and round. Some want it moved to after the 10th to avoid HPCS issues and some want it moved to fall, so its spring time in the southern hemisphere.

Adak

Checking a few forums, this is what I found:

OCF: 100% for not requiring a name change, in order to race

Tech Power Up: 100% for not requiring a name change in order to race

This is a typical quote I read on TPU:
Quote
You will get more participation if every participant gets to keep their folding name. Since they are already folding, it won't affect their standing in that particular team. Plus no need to change anything, its already there. This is the probably the most complained reason thus far (next to the handicap). 


HWC: 100% against it, because they believe it lessens your involvement in the race.

OCN: no poll taken, since Zodac is against it. She states that no poll is needed because obviously the no change needed option would win.

EVGA: for it, percent number is not known

One of the stats systems that devdog51 mentioned, is skewed in favor of the big teams, which the creator plainly state.  I'm not in favor of such a point system. I haven't studied the second one mentioned yet. Let's give it a few days and see if the stats system can be tweaked NOW, instead of after the race.

If we move the race to around either the Spring or Fall Equinox, it will be perfectly balanced for teams in both North and South hemispheres, since that is when the Summer heat and Winter cold on the earth, are perfectly balanced at 12 hours each.

 
Here is what I got on keeping usernames the same for the CC or to change them to the chimp names.

We have Gryphon's formula which can be seen here:

I don't think it's final but that's been met with the least resistance.
 
Team names, I personally think they'll be here to stay, but unsure what the other teams (that aren't helping out at all) have to say. Most likely it'll be put to vote.
 
Date, either May 10th or after Summer is what I've seen, we will see.
 
Adak says EVGA is "for it but no % known". If that's the case, how does he know that EVGA is for it?

I would personally vote for may 15th. It gives us almost a month to work on this system and maybe tweak it a little to Adak's liking. It gives us time for recruitment as well. 

TPU won in Gryphon's latest projection (post 502 linked above). How is that favored against small teams?

csm725OCN


We have Gryphon's formula which can be seen here:

I don't think it's final but that's been met with the least resistance.

Team names, I personally think they'll be here to stay, but unsure what the other teams (that aren't helping out at all) have to say. Most likely it'll be put to vote.

Date, either May 10th or after Summer is what I've seen, we will see.

Adak says EVGA is "for it but no % known". If that's the case, how does he know that EVGA is for it?

He probably looked through the thread to see what the general idea is. The poll that weve taken is right here:
http://www.evga.com/forums/tm.aspx?high=&m=1551383&mpage=1#1551383

csm725OCN


TPU won in Gryphon's latest projection (post 502 linked above). How is that favored against small teams?

Adak

The problem is that Zodac's numbers that this is based on, skew the data - (for instance, in her spreadsheet on EVGA she posted, OCN wins the 2011 CC) - and has been rejected already by EVGA, and other teams.

When you make a "conversion" category to pick the CC overall winner, I  want to know what "conversion" means, in that point system, not asked to maybe count syllables in a name to pick the winner, instead. 

For example, in the column on the left (points only counting), EVGA has a slight edge on OCN. But in the "conversion" column, EVGA has been shrunk to almost half of OCN's numbers. What happened here?

I believe a Jaded monkey (perhaps with a scroll added to it), should be created as a trophy for the flat out points winner in the CC, but that should not be a consideration for the regular Jaded Monkey winner. That part should remain a handicap race, only, perhaps with a scroll added for it, as well:

A Points winner Jaded Monkey, and the current Jaded Monkey trophy, sounds good. An overall combined winner - isn't working yet.

I don't want to see stats systems adopted, with the understanding that they'll be "fixed" later on. I want them to be adopted AFTER they've been fixed! 
 
 

zodac

The conversion was taken from the CC last year - the points made by the CC Name, divided by the points made by the entire team. It was a factor I thought would balance the formula, and we needed some numbers for it. If you have a more efficient/preferred method for the conversion, go ahead.

I like those participation numbers. 

zodac

As for your idea of splitting the trophies, that won't work at all (as opposed to the combined system just needing some tweaks). As I've pointed out, it means only two teams will compete for the points title, and two (maybe three) will compete for the handicap title.

I believe the large teams need some recognition for their tremendous production. I'm surprised you don't like this idea.

Can you quantify a percent increase in growth of a small team, with X number of points increase in the larger teams? That is what you're trying to do, and I'm not sure it's possible to do that, in a fair way. You can do it for races which have already finished, but for unknown input -- no. 

zodac

If you feel that strongly that we shouldn't use any system until it's as close to 100% as possible, then I don't see any way we can hold the CC in May. Gryphon's system, while not yet perfect, seems to be better than last year's. I'm not saying it won't be biased towards a certain team, but we can't really find that out until we run the event.

We have time to still tweak the stats systems, and we should delay accepting a system, until it has been tweaked, and some "probable" CC numbers have been fed into it, to see if it works well with input that is other than the CC from last year. Running "what if" is one of the things a spreadsheet does best, so I don't understand your last comment. 

As for the time squeeze - that's a concern, I agree. What if we announce the starting date for the CC. Now if we use the whole team idea, then there will be plenty of time, since no passkeys need to be sent out, or folding names changed. We continue to tweak the stats, and select the best one by say, April 25th.

That gives the webmaster at least 10 days to get the website tweaked for it, maybe more, depending on the starting date. Big two questions are: -- who is going to set and monitor the web site this year, and how much lead time will he/she need?

We need the answer to the big questions above, before we can set a starting date.

devdog51

The current formula that Zodac likes and that troy created a few pages back. The CCC. Adak is against it because it favors the big teams and needs tweaked some more. As for date thats still going round and round. Some want it moved to after the 10th to avoid HPCS issues and some want it moved to fall, so its spring time in the southern hemisphere.

 
Explanation please.

troy8d

devdog51

The current formula that Zodac likes and that troy created a few pages back. The CCC. Adak is against it because it favors the big teams and needs tweaked some more. As for date thats still going round and round. Some want it moved to after the 10th to avoid HPCS issues and some want it moved to fall, so its spring time in the southern hemisphere.

Explanation please.

The post right above yours is the explanation that he gives. I dont see how it favors big teams over small teams. It may not be perfect but without a paid panel of mathematicians that work on that and only that, are we ever going to achieve a perfect formula? 

troy8d

theGryphon


I don't have the time to argue either. So, please do as you wish.

I do believe they do the same thing, one is just a much simpler approach than the other. 

 
Yeah, that's the thing: they're not doing the same thing. I explained it in my previous post when I implemented geomean and prepared a table. Please see that (#499). It still favors the Points category because that category has a wider spread (range). At the end, it ends up giving larger teams an edge.
 
That is why I put my proposal so that the range in different categories are set equal. Range (maximum minus minimum) is what matters. Once it is equal across categories, you can just add the points from each category to come to a single score that is fair. 
 
To Adak, zodac, and others: The final proposal I made (#502) is fair in the sense that it gives equal importance to all categories. It does not favor any category, as long as the categories are decided on. If anybody thinks that I'm trying to make it so that EVGA wins, they couldn't be more wrong. I'm putting it here, I could care less about who wins. Hell, most probably I won't even participate CC. I just saw Troy's idea, I liked it, and I tried to make it better (balanced). 
 

I read what he wrote and it is not an explanation.  Making a statement like that without any explanation/evidence is meaningless...

theGryphon

Yeah, that's the thing: they're not doing the same thing. I explained it in my previous post when I implemented geomean and prepared a table. Please see that (#499). It still favors the Points category because that category has a wider spread (range). At the end, it ends up giving larger teams an edge.

That is why I put my proposal so that the range in different categories are set equal. Range (maximum minus minimum) is what matters. Once it is equal across categories, you can just add the points from each category to come to a single score that is fair. 

To Adak, zodac, and others: The final proposal I made (#502) is fair in the sense that it gives equal importance to all categories. It does not favor any category, as long as the categories are decided on. If anybody thinks that I'm trying to make it so that EVGA wins, they couldn't be more wrong. I'm putting it here, I could care less about who wins. Hell, most probably I won't even participate CC. I just saw Troy's idea, I liked it, and I tried to make it better (balanced). 

Fair enough.  In all honesty I haven't really had time to look at your modification other than to briefly skim it so I missed out on the nuance of this difference. 

if you mean this post
http://www.evga.com/forums/fb.ashx?m=1550436
 Categorized Chimp Challenge
 
this was similar to one I thought of...
this is where I thought it should go
 
as for the CC names - find out if the teams won't budge off positions
 
1. they be out if CC names stay 
2. they be out if CC names gone
 
if we had to see if you can wrangle something sort of like MLB
 
one year if CC names/ one year isn't
national league rules /american league rules
 
I could settle testing with CC names on for one more year and switch next year
 
 
 
 
 

troy8d


I read what he wrote and it is not an explanation.  Making a statement like that without any explanation/evidence is meaningless...

I understand. Well if we can get everyone else to like it then he'll have to go with it. Zodac already likes it and I think leroy over at OCAU does too. If you want here is the entire conversation:

devdog51

theGryphon

To answer those questions: 
  
1) Honestly, I don't know. I directly used the data for conversion provided by zodac. I assume it's the percentage of Chimp points to Team points, as that's the only way it would make sense. 
2) I don't know who thinks using Points doesn't make sense, and why, but I disagree. That said, I'm not proposing any categories, but just a way to put them together in a balanced way. 
That also said, some don't like Points, some don't like Conversion, some thinks Growth is pretty much the same as Conversion... I don't know what you guys want, count the syllables in our last names and decide the winner that way??? If nobody compromises there won't be a CC, a case which I'm fine with btw... 
3) I don't know what you mean by "handicap", but Conversion is not the same thing as Growth. Growth is percent increase in Chimp name points compared to previous year's CC. A team can have very good Conversion but can suck in Growth (see for example Beavers_Gone_Bananas in the table I attached in my previous post). Another team may well be the opposite (see for example CustomBit Chimps in the table).  
  
My proposal is such that given the categories you (we, whatever) decide on, this is the best way to put them together. This way, for a team to win the CC, they have to do very well in all categories. They don't have to rank the first in any, they just have to "not suck" in any. I think this is great because it would force all teams to do well an all fronts, and not lay back once they believe they're gonna win in one category. 
  
To me, all these three categories I included make sense, but I'm not insisting on any category. You wanna count syllables?, so be it.

Hope this answers it.

zodac

The conversion was taken from the CC last year - the points made by the CC Name, divided by the points made by the entire team. It was a factor I thought would balance the formula, and we needed some numbers for it. If you have a more efficient/preferred method for the conversion, go ahead.

As for your idea of splitting the trophies, that won't work at all (as opposed to the combined system just needing some tweaks). As I've pointed out, it means only two teams will compete for the points title, and two (maybe three) will compete for the handicap title.

If you feel that strongly that we shouldn't use any system until it's as close to 100% as possible, then I don't see any way we can hold the CC in May. Gryphon's system, while not yet perfect, seems to be better than last year's. I'm not saying it won't be biased towards a certain team, but we can't really find that out until we run the event.

 

adak

The problem is that Zodac's numbers that this is based on, skew the data - (for instance, in her spreadsheet on EVGA she posted, OCN wins the 2011 CC) - and has been rejected already by EVGA, and other teams.

When you make a "conversion" category to pick the CC overall winner, I  want to know what "conversion" means, in that point system, not asked to maybe count syllables in a name to pick the winner, instead. 

For example, in the column on the left (points only counting), EVGA has a slight edge on OCN. But in the "conversion" column, EVGA has been shrunk to almost half of OCN's numbers. What happened here?

I believe a Jaded monkey (perhaps with a scroll added to it), should be created as a trophy for the flat out points winner in the CC, but that should not be a consideration for the regular Jaded Monkey winner. That part should remain a handicap race, only, perhaps with a scroll added for it, as well:

A Points winner Jaded Monkey, and the current Jaded Monkey trophy, sounds good. An overall combined winner - isn't working yet.

I don't want to see stats systems adopted, with the understanding that they'll be "fixed" later on. I want them to be adopted AFTER they've been fixed! 

 
And then what I posted above. That is the length of his argument against your system. Maybe he just doesnt like it since its not his?

By the way, just to emphasize something about CC name change: without it,  there won't be a "Conversion" category, so we'll be left with two categories; Points and Growth.
 
Unfortunately, so far I didn't see another performance measure that can't be cheated in.
 

From reading Adak's comments, he does not know what he is talking about.  In fact, it appears his questions are directed entirely at Zodac's format/formula and not the one gryphon and i are discussing.  It seems he is misunderstanding the metrics being used as well as which format is being discussed. 

theGryphon


By the way, just to emphasize something about CC name change: without it,  there won't be a "Conversion" category, so we'll be left with two categories; Points and Growth.

Unfortunately, so far I didn't see another performance measure that can't be cheated in.

 
I think without conversion, CCC is certainly missing something as points and growth alone are not nearly as interesting.  Conversion favored the smaller teams, points favored the larger teams and growth had an equal footing for all teams.

devdog51, please invite Adak to read my posts at #502 and #587, and this one.
 
Just to clarify, I've been using zodac's data from last year's CC (which is not manipulated) only for demonstration purposes. The idea Troy and I have is totally different from zodac's formula, and it is not based on anything, other that trying to be fair and balanced.
 

troy8d


From reading Adak's comments, he does not know what he is talking about.  In fact, it appears his questions are directed entirely at Zodac's format/formula and not the one gryphon and i are discussing.  It seems he is misunderstanding the metrics being used as well as which format is being discussed. 

Well going back in time a little bit, this isnt exactly the first idea that he's rejected simply on what he THINKS it does regardless of what it actually does

devdog51

troy8d


From reading Adak's comments, he does not know what he is talking about.  In fact, it appears his questions are directed entirely at Zodac's format/formula and not the one gryphon and i are discussing.  It seems he is misunderstanding the metrics being used as well as which format is being discussed. 

Well going back in time a little bit, this isnt exactly the first idea that he's rejected simply on what he THINKS it does regardless of what it actually does

 
What is Adak's idea? I don't see it being explained anywhere. If he thinks he has a better idea, why not put it on the table?
 
Well, regardless of what system is gonna be decided on, once CC is finished (assuming it's gonna happen), I will apply the method in #502 and see how it compares, just for giggles...
 

I think he may be rooting for his formula still that was proposed a while back. Im not sure though. Like I said before, maybe I can persuade the other active teams and we can have a majority to overrule him.

I like 502 the most, but of course it insinuates that we have team names.

Well team names might be something that we have to put up with. I would rather see them gone but in the spirit of cooperation, Ill keep the names if it means that we implement the new system.